KARS-related diseases: progressive leukoencephalopathy with brainstem and spinal cord calcifications as new phenotype and a review of literature

KARS encodes lysyl- transfer ribonucleic acid (tRNA) synthetase, which catalyzes the aminoacylation of tRNA-Lys in the cytoplasm and mitochondria [1]. Mitochondrial and cytoplasmic aminoacyl–tRNA synthetases (aaRSs) are encoded by distinct nuclear genes, with the exception of KARS and GARS (glycyl–tRNA synthetase) which are present in both cellular compartments [2, 3].

Mutations in aaRSs genes have been linked to a growing number of neurological and systemic disorders with heterogeneous phenotype. Eleven families/sporadic patients and 18 different mutations in KARS have been reported to date. Phenotype is heterogeneous ranging from early onset encephalopathy [4‐7] to isolated peripheral neuropathy [8] or nonsyndromic hearing impairment [9]. Recently, late onset leukoencephalopathy [10] and cardiomyopathy [11, 12] have been reported.

A progressive leukoencephalopathy with brainstem and spinal cord calcifications was previously described as a distinct entity in a singleton patient [13] and in two siblings [14]. We report the clinical, biochemical and molecular findings of 2 unreported patients presenting a similar clinical and radiological picture. Genetic analysis performed on them and in the patient previously described [13] revealed the presence of biallelic mutations in KARS in all three subjects. Review of KARS mutant patients published to date will also be discussed.

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2000. Written informed consent was obtained from all individuals or caregivers.

Mutations in different aaRSs have been associated with an increasing number of phenotypes [2, 3]. Encephalopathy is the most common phenotype, but other extra neurological symptoms have been reported: sideroblastic anemia (YARS2 [17, 18]), cardiomyopathy and myopathy (YARS2 [19], GARS [20], KARS [11, 12], tubulopathy (SARS2 [21]), ovarian failure (AARS2 [22], HARS2 [23], LARS2 [24]), hepatopathy (FARS2 [25], EARS2 [26]) and hearing loss (HARS2 [23], LARS2 [25], KARS [9]).

Up to now, 11 families/sporadic patients and 18 mutations in KARS have been reported (Tables 1 and 2).

KARS impairment was first linked to peripheral neuropathy [8] in one patient (Pt 1) presenting with Charcot-Marie-Tooth neuropathy, developmental delay, self-abusive behavior, dysmorphic features, and vestibular Schwannoma. Compound heterozygous p.Leu133His and p.Tyr173SerfsX7 variants were identified. Functional analyses revealed that these two mutations severely affect enzyme activity.

Autosomal recessive nonsyndromic hearing loss was the second phenotype reported [9]. In affected individuals homozygosity for missense mutations (p.Asp377Asn or p.Tyr173His) in KARS was identified. (Pts 2–14/Fam. 2–4). Both variants were predicted to be damaging by multiple bioinformatics tools. The first case of KARS mutations associated with suspected mitochondrial disease was reported in 2013 [4]. The authors analyzed by exome sequencing a series of 102 patients with clinical and biochemical findings suggestive for mitochondrial disorders and identified compound heterozygous KARS mutations (p.Thr587Met; p.Pro228Leu) in a patient affected by psychomotor delay, hearing loss, ophthalmoplegia, dystonia and elevated CSF lactate level (Pt 15/Fam. 5). MRC activity on tissue was not investigated. No experimental proof was reported but, given the predicted severity of the mutations at highly conserved residues, the authors concluded that the observed mutations were likely the genetic cause of patient’s phenotype.

A more severe phenotype was reported in two 2 siblings (Pts 16–17/Fam. 6) with early onset visual impairment, progressive microcephaly, developmental delay, seizures and very subtle deep white matter loss on MRI [5]. The patients harbored compound heterozygous mutations (p.Arg466Trp; p.Glu553Lys) within a highly conserved region of the catalytic domain. A similar clinical presentation was reported in a patient who harbored a p.Ala57Pro missense change and a 7601-base pair deletion, encompassing the first three exons of the mitochondrial isoform of KARS (Pt 18/Fam. 7) [6]. Cardiac involvement associated with a deficiency of MRC complexes I and IV has been reported in two patients (Pts 19–20/Fam. 8–9), who carried novel biallelic KARS mutations [11, 12]. The first one presented a childhood-onset hypertrophic cardiomyopathy associated with seizures, developmental delay, in a patient harboring compound heterozygous p.Val476Asp and p.Ile346Thr mutations [11] while hypertrophic cardiomyopathy was the clinical hallmark in the second, a 14 years old patient with mild myopathic signs and cognitive disability (in spite of normal brain MRI) associated with p.Leu378His and p.Pro418Arg [12]. In both patients lactic acidosis was detected. In the first case, the mitochondrial enzyme defects were rescued by cDNA complementation with mitochondrial KARS, but not cytosolic form [11]. More recently, two mutations (p.Arg505His; p.Pro533Ser) have been reported in two siblings affected by early onset hearing loss, progressive cognitive impairment and psychiatric symptoms with onset in adulthood associated with leukoencephalopathy: brain MRI showed symmetrical confluent abnormalities in the frontal, periventricular white matter and in the corpus callosum [10]. Functional studies showed that both mutations decreased tRNA aminoacylation while p.Arg505His changed the secondary structure of KARS, leading to protein aggregation. Finally, KARS mutations (p.Ala526Val; p.Phe489Cys) were reported in two sisters affected by developmental delay, microcephaly, seizures, and sensorineural hearing loss; calcifications of left occipitoparietal junction were reported in one case (Pt 23). MRC enzymes activity in muscle biopsy was normal, lactate level was not available (Pt 24) [7].

In patient B, the severity of the phenotype, the clinical onset related to febrile illness and the presence of lactic acidosis suggested a mitochondrial disease that was directly investigated in spite of normal MRC and PDH activities in fibroblasts.

In patient C, clinical picture and lactic acidosis claimed the idea of a mitochondrial disorder as already suggested also in patient A, supported by mild lactate elevation at spectroscopy study. In KARS-mutant cases reported to date, mitochondrial disease was suspected and confirmed by biochemical diagnosis only in patients with cardiomyopathy (Pt 19 and Pt 20); elevated lactate level in CSF was detected in Pt 15 but biochemical studies on tissue were not performed. In other cases, metabolic analyses for mitochondrial disease were unremarkable or not performed. In 5 of 26 KARS patients reported to date, including our cases, both lactate level and biochemical studies (in different tissues) were performed and only in patients with cardiac involvement elevated lactate level corresponded to reduced MRC activity. Nevertheless the lack and heterogeneity of laboratory data does not permitted this phenotypic variability explanation. The MRI findings were similar in all three patients and characterized by progressive diffuse leukoencephalopathy and calcifications extending in cerebral, brainstem and cerebellar WM, with spinal cord involvement. Specifically, at the early stages of the disease the signal abnormalities were observed in the deep cerebellar WM and in the centrum semiovale. Progressively, an extensive diffuse WM involvement, including U fibers, posterior arm of the internal capsules, external capsules, thalami, cerebellar peduncles and brainstem, with selective bilateral symmetric involvement of the bulbar pyramids and lateral bulbar regions resembling the pattern of mitochondrial diseases were observed. The supratentorial WM involvement was characterized by uniform slight hyperintensity on T2WI, interrupted by marked foci of hypointensities due to calcifications. This appearance seems to be due to a demyelinating process an assumption supported also by the evidence of demyelination of the proximal intracisternal portion of the V cranial nerves (Fig. 1g).

Cerebral calcifications have a distinct pattern with initial involvement of deep cerebellar and cerebral periventricular WM and progressive extension to the thalami and internal capsules, in which a peculiar “boomerang appearance” was present. Calcifications were evident even in the initial phases of the disease and might not be a dystrophic epiphenomenon and so a secondary and aspecific event, but an intrinsic feature of the disorder. In the spine they were present in all 3 patients, even if with different severity, and were characterized by a peculiar bilateral and symmetrical distribution in the anterior horns, both extensively (Patient A) or spot-like (Patients B and C). On MRI, bilateral abnormal signal intensity in the lateral columns was also associated.

Patient B dysplayed a more severe cerebral atrophy and grey matter involvement (basal ganglia and cortex) but he underwent the first MRI later in life compared to the other two. As disease progressed, the radiological picture evolved toward a progressive cerebral atrophy in 2 patients (A and B).

WM involvement has been previously reported in few KARS-mutant patients but with a less severe pattern and restricted to supratentorial regions (Pt 16, Pt 17, Pt 21, Pt 22). It is interesting to note that the presence of cerebral WM and spinal cord signal abnormalities is a pretty rare association of neuroradiological features and it is typically observed in other aaRSs deficiencies, notably in DARS and DARS2 related leukodystrophies [27, 28]. It is also a quite common finding in Iron-sulfur cluster related leukoencephalopathies, particularly those caused by GLRX5 [29], ISCA2 [30], or IBA57 [31] mutations. The association of cerebral WM abnormalities with spinal cord involvement should prompt to consider aaRSs-related diseases and particularly KARS mutations when calcifications are observed.

The extremely heterogeneous clinical presentation associated with KARS mutations is peculiar in the field of aaRS-related diseases which are usually characterized by strict genotype-phenotype associations, although a definite explanation of the molecular mechanisms underpinning this observation is still missing. Few examples of different phenotypes caused by mutations in the same aaRS gene have been reported (e.g. AARS2 associated with either cardiomyopathy or leukoencephalopathy and ovarian failure [22, 32]). Differences in mode of inheritance and type of mutation cannot easily explain the variable clinical presentations since all the reported cases showed an autosomal recessive transmission of missense mutations. Only the patient described by Joshi et al. (Pt 18/Fam. 7) carried a large deletion, acting as a null allele, together with a missense mutation which disrupts the mitochondrial targeting signal, thus potentially affecting solely the mitochondrial isoform of KARS. All the other patients, irrespective of any evidence of mitochondrial dysfunction, harbored KARS variants which are predicted to strike both mitochondrial and cytosolic KARS isoforms. An effect of the affected functional domains was initially suggested, since the first mutations responsible for the neuropathic phenotype hit the anticodon domain whereas hearing loss associated mutations could be in the catalytic domain. However this hypothesis was not confirmed in the following reports and in the present review of all the KARS mutant patients. For instance, the mutations found in our patients, with an overlapping phenotype, are scattered throughout the gene (from amino acid 127 to 505) and affect either the anticodon-binding or the catalytic domain. The few functional studies which have been performed indicated that various functions/properties of KARS (e.g. tRNA aminoacylation, secondary structure) may be affected by different mutations. Nevertheless, no genotype/phenotype correlation was evident, even considering the residual enzymatic activity of the different mutant forms. Nevertheless a mutation specific effect cannot be excluded, since all the identified KARS mutations were reported in single cases/families; for instance, the cardiac phenotype in AARS2 mutant patients seems to be strictly linked to the presence of a specific amino acid change. The only KARS mutation presents in two unrelated families was the p.Arg505His, identified in homozygosity in patient A and in compound heterozygosity with p.Pro533Ser in Pts 21–22; all these three individuals were characterized by leukodystrophy and hearing problems but the MRI features were not identical and other clinical symptoms were different (e.g. visual impairment and spastic tetraparesis were observed in patient A but not in the two siblings). Obviously, the partially different genotype may account for the phenotype diversities.

With our report we define the molecular basis of the previously described Leukoencephalopathy with Brainstem and Spinal cord Calcification, that we propose to call LBSC similarly to DARS and DARS2 related leukodystrophies, widening the spectrum of KARS related disorders, particularly in childhood onset disease suggestive for mitochondrial impairment. The review of previous cases does not suggest a strict and univocal genotype/phenotype correlation for this highly heterogeneous entity.

Moreover, our cases confirm the usefulness of search for common brain and spine MR imaging pattern and of broad genetic screening, in syndromes clinically resembling mitochondrial disorders in spite of normal biochemical assay.